Many approaches to robots that resemble animals with skeletons are being actively developed. Most of these robots are constructed using so-called “hard” body plans; that is, a rigid (usually metal) skeleton, electrical or hydraulic actuation, electromechanical control, sensing, and feedback. These robots are successful at the tasks for which they were designed (e.g., heavy manufacturing in controlled environments.) However, these robots have severe limitations when faced with more demanding tasks (for example, stable motility in demanding environments): tracks and wheels perform less well than legs and hooves in many circumstances.
Another class of robots—those based on animals without skeletons—are much less explored, for a number of reasons: i) there is a supposition that “marine-like” organisms (squid) will not operate without the buoyant support of water; ii) the materials and components necessary to make these systems are often not available; iii) the major types of actuation used in them (for example, hydrostats) are virtually unused in conventional robotics. These systems are intrinsically very different in their capabilities and potential uses than hard-bodied systems. While they will (at least early in their development) be slower than hard-bodied systems, they will also be more stable and better able to move through constrained spaces (cracks, rubble), lighter, and less expensive.
Robots, or robotic actuators, which can be described as “soft” are most easily classified by the materials used in their manufacture and their methods of actuation. The field of soft robotic actuation began with work by Kuhn et al in 1950. Their work focused on the reversible change in the coiling and uncoiling of a polymeric material dependent on the pH of the surrounding medium. They used this to successfully raise and lower a weight, thus showing proof of principle for the use of soft materials in robotic actuation. Hamlen et al expanded upon this idea in 1965 and showed that polymeric materials can be made to contract electrolytically. These two developments set the scene for future work using the swelling of polymeric gels and electronic control of dielectric-based actuators. Otake et al have demonstrated the use of electro-active polymers in the manufacture of starfish-shaped robotic actuators. Pneumatically-driven soft actuators based on pressurization of sealed chambers fabricated from extensible polymers were first reported by Suzumori et al in 1991. This type of actuation has been used on the millimeter scale to fabricate grippers, tentacles, and other related devices including pneumatic balloons.
Pneumatic soft robotic actuators can be manufactured using inextensible materials, which rely on architectures such as bellows. McKibben actuators, also known as pneumatic artificial muscles (PAMs), rely on the inflation of a bladder constrained within a woven sheath which is inextensible in the axis of actuation. The resultant deformation leads to radial expansion and axial contraction; the force that can be applied is proportional to the applied pressure. Related actuators are called pleated pneumatic artificial muscles.
There are “soft” robotic actuators such as shape memory alloys which have been used by Sugiyama et al both as the actuation method and as the main structural component in robots which can both crawl and jump. Another approach, which can be described as “soft” uses a combination of traditional robotic elements (an electric motor) and soft polymeric linkages based on Shape Deposition Manufacturing (SDM). This technique is a combination of 3D printing and milling. An example of a composite of traditional robotics with soft elements has been used with great success in developing robotic grippers comprising soft fingers to improve the speed and efficiency of soft fruit packing in New Zealand.
Additional capabilities for soft robotics are desired.